A 25 km streak of CF3SF5 was released on an isopycnal surface approximately 1100 m deep, and 150 m above the bottom, along the continental slope of the northern Gulf of Mexico, to study stirring and mixing of a passive tracer. The location and depth of the release were near those of the deep hydrocarbon plume resulting from the 2010 Deepwater Horizon oil well rupture. The tracer was sampled between 5 and 12 days after release, and again 4 and 12 months after release. The tracer moved along the slope at first but gradually moved into the interior of the Gulf. Diapycnal spreading of the patch during the first 4 months was much faster than it was between 4 and 12 months, indicating that mixing was greatly enhanced over the slope. The rate of lateral homogenization of the tracer was much greater than observed in similar experiments in the open ocean, again possibly enhanced near the slope. Maximum concentrations found in the surveys had fallen by factors of 104, 107, and 108, at 1 week, 4 months, and 12 months, respectively, compared with those estimated for the initial tracer streak. A regional ocean model was used to simulate the tracer field and help interpret its dispersion and temporal evolution. Model-data comparisons show that the model simulation was able to replicate statistics of the observed tracer distribution that would be important in assessing the impact of oil releases in the middepth Gulf.